USGS “iCoast – Did the Coast Change?” Project: Crowd-Tagging Aerial Photographs to Improve Coastal Change Prediction Models

Abstract:

The U.S. Geological Survey (USGS) has been acquiring oblique aerial photographs of the coast before and after major storms since 1995 and has amassed a database of over 140,000 photographs of the Gulf, Atlantic, and Pacific coasts. USGS coastal scientists use these photographs to document and characterize coastal change caused by storms. The images can also be used to evaluate the accuracy of predictive models of coastal erosion. However, the USGS does not have the personnel to manually analyze all of the photographs taken after a storm. Also, computers cannot yet automatically identify damages and geomorphic changes to the coast from the oblique aerial photographs. There is a high public interest in accessing the limited number of pre- and post-storm photographic pairs the USGS is currently able to share. Recent federal policies that encourage open data and open innovation initiatives have resulted in many federal agencies developing new ways of using citizen science and crowdsourcing techniques to share data and collaborate with the public to accomplish large tasks. The USGS launched a crowdsourcing application in June 2014 called “iCoast – Did the Coast Change?” (http://coastal.er.usgs.gov/icoast) to allow citizens to help USGS scientists identify changes to the coast by comparing USGS aerial photographs taken before and after storms, and then selecting pre-defined tags like “dune scarp” and “sand on road.” The tags are accompanied by text definitions and pictorial examples of these coastal morphology terms and serve to informally and passively educate users about coastal hazards. The iCoast application facilitates greater citizen awareness of coastal change and is an educational resource for teachers and students interested in learning about coastal vulnerability. We expect that the citizen observations from iCoast will assist with probabilistic model development to produce more accurate predictions of coastal vulnerability.